The 65-billion-parameter model is available under Apache 2.0. Ricursive reports a strong coding benchmark score, but results vary across tests.

Ricursive Intelligence has released Ripple-65B Preview, an open-weights AI model that the startup says leads U.S. open models on Terminal-Bench

パラメーター数650億のモデルをApache 2.0ライセンスで公開。Ricursiveはコーディングベンチマークで高得点を報告していますが、テストによって結果は異なります。

Ripple-65B Preview is licensed under Apache 2.0 and scores 80.5 on Terminal-Bench 2.1. Details about access and test conditions were not provided.

Ricursive's Ripple-65B Preview is a 65B open-weight model under Apache 2.0 that it says tops U.S. open models at agentic coding. What it means for dev teams.

Ricursive Intelligence has officially announced the release of Ripple-65B Preview, a high-performance open-source weight model designed to advance agent-based programming and large-scale data processing. Developed with a…

Ricursive Intelligence’s Ripple-65B Preview is a free 65B open-weight model it says beats Reflection Beam and Meta Muse Spark at coding, built for $32m.

We develop frontier AI methods to reinvent chip development, collapsing traditional design timelines, closing the loop between AI and the hardware that fuels it, and enabling a Cambrian explosion of…

Independent rankings of software and IT services by category and country, daily tech and business insights, and reviews from real users. Free for buyers.

MarkTechPost

Ricursive Intelligence (à ne pas confondre avec Recursive, à l'origine de Flint) publie la preview de Ripple-65B, un modèle dense de 65 milliards de paramètres. D'après ses benchmarks, il rivalise ave

AI企業が新しい言語モデルを出した、というだけなら今では珍しくありません。ですがRicursive Intelligenceの発表は少し毛色が違います。彼らは…

Frontier code agents need to run long tasks in a terminal without losing track, and pretraining them is usually billed in hundreds of millions. That is the gap Ricursive says it built Ripple-65B Preview to close, a 65-billion-parameter model pretrained on 12 trillion tokens with support for up to 1M context length. Ripple-65B Preview is a dense Transformer, an architecture that activates all parameters for every token, and Ricursive says the general-purpose variant of the proprietary model that powers its AI chip design platform sets a new U.S. open-weight state of the art at 80.52 on TerminalBench 2.1, a benchmark that scores agents on executing real coding tasks inside a Linux terminal. That score beats significantly larger U.S. open models on the same benchmark. Ricursive puts Reflection Beam 501B at 80.1, Meta Muse Spark Open 816B at 76.4, Thinking Machines Inkling 975B at 63.8 and NVIDIA Nemotron 3 Ultra 550B at 56.4, according to the company’s release. The bet is that a smaller dense model can win on latency and local deployability where sparse models need more memory. Ripple-65B Preview is built to fit devices like NVIDIA DGX Spark, which Ricursive notes has high FLOPs with low memory bandwidth versus...